General Triangle Drafting: A faster asynchronous drafting format

Jason Waddell

Administrator
Staff member
You have my noisy neighbors to thank for this one. After they blasted music at 1 in the morning (a very late hour for new parents), I couldn't fall back to sleep. It wasn't just the noise. A thought was eating away at me.

'There has to be a better way to do asynchronous online 8-player drafts'

We've done a very literal translation of booster drafting on the forums before. Crack a pack of 15 cards, remove one, post the remaining contents to the next drafter. Open a pack of 14 cards, remove one, post the remaining contents to the next drafter.

Rinse and repeat. 42 times.

42 times that each player has to pass a pack. The process was such a drag. So many bottlenecks waiting on players to catch up on their picks. Even if you pass a couple packs a day, the whole thing takes a good two-to-three weeks. And the ends of packs are particularly brutal. Open a 4-card pack, take the only card that is relevant to your deck (if there even is one), pass it on.

I started to think if there was a way to work backwards. Those tedious ends of packs, when you have 1 card left, 2 cards left, 3 cards left, was there a way we could get through them more quickly?

The count began.

1, 2, 3, 4, 5.

1 + 2 + 3 + 4 + 5 = 15. That's interesting. What if we keep counting for each player.

1 + 2 + 3 + 4 + 5 + 6 + 7 + 8. Thirty-six.

Wait a minute. Are you thinking what I'm thinking? Do it. Go one step further.

1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 + 9 = Fourty-five. It's beautiful. Three packs of 15. One-eighth of a cube.

From there the format pretty much writes itself.


How it works:

1) Each player starts with one triangle. The triangle consists of 9 rows.

A row of 9 cards.
A row of 8 cards.
A row of 7 cards.
A row of 6 cards.
A row of 5 cards.
A row of 4 cards.
A row of 3 cards.
A row of 2 cards.
A row of 1 card.

2) Look at your whole triangle. Draft 1 card from each row of the triangle.

3) Pass the remaining triangle (now 8-rows) to the next drafter. Repeat until all cards are drafted.

I used a quick bit of R code to generate triangles from my cube. Let's take a look at one to help visualize the process.




















If you're anything like me, this will set your brain alight. There are so many directions you could take this. Gut reactions:

Naya Aggro


Green Ramp (no second color commitment yet)


Jeskai Tempo


Blue-White Control


...and we didn't even touch Dark Confidant!

I could spend a whole lunch break just chewing over these juicy options. Even within an archetype the answers aren't obvious. Does my tempo deck want Daze of Treasure cruise?

The best part of it all?

Eight pack passes per player. 8 versus 42! We could finish one of these drafts in what, two, three days? I am so excited!

What will the meta be like? Do you commit hard with your first pack or keep your options open in case you get cut off? How do you adapt to the delayed signaling?

And what would you take from that triangle?
 
So you pick from:
9/8/7/6/5/4/3/2/1
8/7/6/5/4/3/2/1
7/6/5/4/3/2/1
...
1

In triangle draft, your average pick is from 3.67 cards. In a regular 3x15 draft, your average pick is from 8 cards. Doesn't that make any type of microarchetype drafting way, way harder, especially with the lack of signals?
 

Jason Waddell

Administrator
Staff member
So you pick from:
9/8/7/6/5/4/3/2/1
8/7/6/5/4/3/2/1
7/6/5/4/3/2/1
...
1

In triangle draft, your average pick is from 3.67 cards. In a regular 3x15 draft, your average pick is from 8 cards. Doesn't that make any type of microarchetype drafting way, way harder, especially with the lack of signals?

This is a good question. Something to keep an eye on when we test it. It's possible that the format sucks after the first triangle(s).
 
From what I can see of how it would function, it's more like a hybrid between sealed and draft. That's not necessarily a bad thing, just a different deck building style.

Unless I did my math wrong, your average pick in traingle drafting is from amongst 18.3 cards per round, not 3.67. Maybe it's 3.67 per "pack" that you see, but that doesn't really matter because you are drafting 9 cards in one go, 8 cards, etc. etc. That 18.3 is why I suspect it will feel a lot more like sealed than drafting. Basically sharing a big sealed pool between the 8 drafters in an organized fashion.
 
  • Like
Reactions: dbs

Jason Waddell

Administrator
Staff member
I don't think the average is the most informative metric. Probably total number of cards seen is more relevant, and there you see about 110 more cards in a booster draft, if my quick mental math is correct. But even then it doesn't really capture the entire dynamic.

It's a complicated question, but in general every drafting mechanism has different output. Sealed, booster draft, grid, rotisserie, etc all produce different decks.
 
The average would I think inform how much flexibility/insight you would have on making archetype decisions. The total number being lower (by 111 to be exact) adds to my conviction that the decks will feel closer to sealed (if sealed is on the other end of the draft-sealed spectrum)
Sealed: 90 cards seen, avg 90 cards seen at a time -> higher flexibility of archetype choice, lower overall consistency of card power etc
triangle: 165 cards seen, avg 18.3 cards seen at a time
draft: 276 cards seen, avg 8 cards seen at a time ->lower flexibility of archetype choice, higher consistency of final pool
 

Jason Waddell

Administrator
Staff member
I think the point japan brings up is quite valid, especially when you consider that each player will get 9 'automatic' one- card row picks. That's 20 percent of the drafted cards.

That said, cube decks have historically not been short on playables.

I think the answer, as always, is to just test it. Worst case we change it to something like five rows of 9 or four rows of 11 and keep the core mechanics in tact.
 

Jason Waddell

Administrator
Staff member
Anybody interested in doing a draft starting tomorrow or so? So far we have myself, James, Dom, Chris MTGO and Kirblinx
 
Top